International Journal of Public Opinion Research Vol. 25 No. 4 2013
ß The Author 2013. Published by Oxford University Press on behalf of The World Association
for Public Opinion Research. All rights reserved.
doi:10.1093/ijpor/edt016 Advance Access publication 19 June 2013
Focusing on Differences? Contextual Conditions and
Anti-immigrant attitudes’ Effects on Support for
Turkey’s EU Membership
Amsterdam School of Communication Research, University of Amsterdam, Amsterdam,
The Netherlands
Abstract
A key predictor of support for Turkey’s EU membership is citizens’ attitudes toward
immigrants. How anti-immigrant attitudes relate to support is not obvious and has
remained unclear thus far. We argue that the presence of immigrants in a country
reduces support for Turkey’s membership. Second, the stronger this presence, the
stronger the effect of anti-immigrant attitudes on support. Third, the more salient
immigration issues in a country, the stronger the effect of anti-immigrant attitudes on
support. Using cross sectional survey data (N ¼ 26,344) from 20 EU member states,
we find support for these claims. Our findings provide a prime example of how
politicians and media can influence citizens’ thinking about specific issues, and, in
doing so, can influence public opinion.
Turkey’s road to EU membership seems to be never-ending. The road may
not come to an end as long as support for Turkey’s membership among EU
citizens is low. This is because several European leaders have committed to
holding a referendum on Turkey’s membership. It is, therefore, important to
investigate what determines public support for Turkey’s accession. Why is it
so low?
Research on public support for EU enlargement with Turkey has mainly
concentrated on individual-level factors. Recent studies focus on citizens’ economic and political attitudes (e.g., Gerhards & Hans, 2011; Karp & Bowler,
All correspondence concerning this article should be addressed to Rachid Azrout, Amsterdam School of
Communication Research, University of Amsterdam, Kloveniersburgwal 48, 1012 CX, Amsterdam, The
Netherlands. Email: [email protected]
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Rachid Azrout, Joost H.P. van Spanje and
Claes H. de Vreese
FOCUSING ON DIFFERENCES?
481
Theory
How can we explain differences in the level of support for Turkey’s EU
membership? To answer this question, we follow research by
McLaren (2007). In that study, McLaren argues that it is the proportion of
a country’s population that is of Turkish origin that explains the degree of
support.
Following Allport’s Intergroup Contact Theory (Allport, 1954), which
states that intergroup contact would reduce prejudice, it is plausible to
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
2006). De Vreese, Boomgaarden, and Semetko (2008) compare the strength of
these predictors and show that anti-immigrant attitudes are relatively strong.
Work by Azrout, Van Spanje, and De Vreese (2011) confirms this and shows
that the effect of anti-immigrant attitudes is driven by citizens’ categorizing
others as an out-group in combination with their negative bias toward outgroups.
Support for Turkey’s accession varies systematically not only between
individuals but also between countries (McLaren, 2007). This leads one to
suspect that aspects related to individuals’ political and media context play a
role here. The impact of contextual factors on public support for Turkey’s EU
entrance has, however, remained largely unexplored. This is perhaps surprising, as contextual factors explain cross-country differences in levels of, e.g.,
support for EU integration (e.g., Anderson & Reichert, 1996) and crosscountry differences in, e.g., effects of individual-level predictors on EU support (e.g., Garry & Tilley, 2009).
Notable exceptions to the rule that contextual effects have not been studied with regard to Turkey’s membership include Jones and Van der Bijl
(2004) and McLaren (2007). Jones and Van der Bijl (2004) show that trade
with and distance to the candidate country matter for levels of support for
entrance of that candidate country. McLaren (2007) demonstrates that the
effect of anti-immigrant attitudes on support for Turkey’s membership substantially varies by country. In this article, we build on these studies and
extend them, providing a comprehensive account of the role of context in
public support for Turkey’s EU access. We contribute to the literature by
examining the effect of various contextual factors on support for Turkey’s EU
membership. Applying the Negative Contact Hypothesis (e.g., Putnam, 2007;
Taylor, 1998), we aim to examine how the presence of and experience with
out-groups (citizens of Turkish origin as well as Muslims) explain differences
in levels of support for Turkey’s EU entry. Also, because we focus on outgroups at the contextual level, we examine the question how the interaction
with anti-immigrant attitudes, as an indicator for perceptions of out-groups at
the individual level, affects support for Turkey’s accession.
482
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
H1: The greater the share of the population from Turkish descent in a country,
the greater the opposition against Turkey’s membership in that country.
However, an issue with the concentration of Turks as a contextual explanation
is the question whether Turkish immigrants come under the rubric of
‘‘Turks’’ or under the rubric of ‘‘Muslims.’’ It has been argued that Turks
are increasingly stereotyped as Muslims (e.g., Marranci, 2004; Poynting &
Mason, 2007). However, Turks are not the only Muslims in the world and
Turkish immigrants are also not the only Muslims living in EU member
states. But Muslims are perhaps easier recognized as a group than are
Turks, for instance by the head scarf women wear. This means that if the
contextual effect is about high concentrations of a discernible group, then
perhaps the group we should focus on is the more easily recognizable group
of Muslims, rather than on Turks.
Van Spanje, Azrout, and De Vreese (2010) show that the dominant
religion among the population of applicant countries is important in explaining
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
assume that the presence of people of Turkish descent in a country increases
support for the Turkey’s candidacy. This is because through contact with
Turks, other citizens in the country become familiar with them, getting a
more favorable attitude toward them in the process. However, intergroup contact is likely to reduce intergroup prejudice under four conditions: Shared
common goals, inter-group cooperation, equal status between the groups,
and support by the authorities (Allport, 1954; Pettigrew & Tropp, 2006;
Taylor, 1998). However, it is argued that under unfavorable conditions, contact
may increase prejudice and intergroup tension (Amir, 1976; Forbes, 1997).
Sniderman, Hagendoorn, and Prior (2004) show that direct contact between
native citizens and immigrants under such conditions is scarce in contemporary
Europe. According to Paolini, Harwood, and Rubin (2010), negative contact
increases the importance of group categorization. When group categorization is
more important, conflict may arise through competition over limited economic
resources, cultural differences, or political struggle (Blalock, 1967; Putnam,
2007; Taylor, 1998). In line with this, several authors find that high concentrations of immigrants may lead to feelings of hostility toward these immigrants among natives (e.g., Bobo, 1988; Giles & Hertz, 1994; Oliver &
Mendelberg, 2000; Quillian, 1995; Taylor, 1998; Valenty & Sylvia, 2004).
Whether from an economic, cultural, or political (perceived) conflict,
McLaren argues that the presence of Turkish immigrants may lead to feelings
of hostility toward these immigrants, and that ‘‘such hostility is likely to carry
over into feelings about whether a country ought to be able to join the
European Union, and that higher concentrations of Turkish immigrants in a
member state will mean individuals in that member state have a lower probability of supporting Turkish candidacy’’ (McLaren, 2007, p. 259). Building
on her work, we first hypothesize:
FOCUSING ON DIFFERENCES?
483
H2: The greater the share of the population in a country that is Muslim, the
greater the opposition against Turkey’s membership in that country.
There is, however, another issue with the number of Turks as a contextual
predictor, and the same issue with the number of Muslims. There are few
Turkish migrants and Muslims (in the EU the proportion Turkish immigrants
does not exceed 3% and the proportion Muslims 6%), and they tend to be so
geographically concentrated that most EU citizens do not have everyday contact with them (Sniderman et al., 2004). As the Negative Contact Hypothesis
suggests that it is physical proximity that leads to hostility (Putnam, 2007,
p. 142), we should perhaps not expect any strong effects.
We propose to extend the argument and argue that direct contact with
Turkish migrants or Muslims is not required. Conflict theory states that
physical proximity of ‘‘others’’ leads to higher levels of distrust of these
others. However, physical proximity may not even be necessary if there is
mediated proximity. We propose that such mediated proximity is the degree to
which the immigration issue is important in a polity. We argue that the more
important the immigration issue, the more attention is given to the issue in
interpersonal communication, and in political and media debates. As a result
of a higher salience of the immigration issue in interpersonal, political, and
media communications, citizens are reminded of the presence of immigrants or
Muslims time and again. Through this, citizens are likely to engage in more
elaboration about the presence of immigrants, which acts as a substitute of
direct contact.
Obviously, importance of the immigration issue is more likely in a country
that has seen higher levels of immigration than in a country with a relative low
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
support for EU membership. By means of an experiment, they show that knowledge about the fact that about half of the population of Bosnia-Herzegovina is
Muslim leads to significantly and substantially less support for Bosnia’s
membership among British, Danish, and German citizens. Van Spanje et al.
argue that fear of Islam is what drives this effect, as they also show that this
effect is moderated by the degree respondents fear immigrants’ religious
practices.
From this, we can follow the same argument as previously, but replace
‘‘Turkish immigrants’’ and ‘‘Turkey’’ with ‘‘Muslims’’ and ‘‘a Muslim country,’’ respectively. We thus argue that high concentrations of Muslims are
likely to be associated with strong feelings of hostility toward Muslims, and
that ‘‘such hostility is likely to carry over into feelings about whether a Muslim
country ought to be able to join the EU, and that higher concentrations of
Muslims in a member state will mean individuals in that member state have a
lower probability of supporting a Muslim country’s candidacy [authors’ alterations of the original quote in italics]’’ (McLaren, 2007, p. 259). We thus
hypothesize:
484
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
H3: The more important the immigration issue in a country, the greater the
opposition against Turkey’s membership in that country.
Given that our aforementioned hypotheses are based on the Negative Contact
Hypothesis, and that we expect citizens to define Turkish immigrants as an
out-group, we should also look at how these contextual factors relate to the
effect of anti-immigrant attitudes on support for Turkey’s accession. Antiimmigrant attitudes are among the strongest predictors of EU attitudes
(e.g., De Vreese et al., 2008; McLaren, 2002) and its effect is driven by
whether individuals are inclined to categorize the world in terms of ingroup and out-groups (Azrout et al., 2011; De Vreese & Boomgaarden,
2005). Therefore, what would be the effect of the presence of a (perceived)
out-group? McLaren finds evidence that the strength of the effect (also) depends on the number of Turkish immigrants (2007, p. 270), but how can we
understand why anti-immigrant attitudes are a more important predictor in
countries where a larger share of the population is of Turkish descent than in
countries with relatively fewer Turkish immigrants?
Azrout et al. (2011) propose a model of how the effect of anti-immigrant
attitudes on support for Turkey’s potential accession might work. Building on
Social Identity Theory, they argue that the effect can be understood in three
steps: First, anti-immigrant attitudes are not so much interesting for researchers as an attitude, but the collection of several evaluations of immigrants
are to be seen as a measure of a general tendency to categorize others in an ingroup and an out-group. Second, when someone with strong anti-immigrant
attitudes is confronted with a specific issue such as EU enlargement, his or
her general tendency to categorize is used to ‘‘frame’’ the issue of the country’s entrance in terms of in-group or out-group. Third, if the candidate
country is classified as an out-group member, either on economic or cultural
grounds, the EU citizen will be negatively biased toward that ‘‘other’’, either
through in-group favoritism or out-group rejection.
For individuals ‘‘to frame’’ is described in the literature as the process of
selecting certain aspects of reality to guide individuals’ processing of information (e.g., Scheufele, 1999; Scheufele & Tewksbury, 2007). By selecting
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
number of immigrants. Yet, the degree of importance of the immigration issue
is partly independent of the share of (Turkish) immigrants and the share of
Muslims. In particular, public opinion on immigration tends to be fickle,
whereas immigrant shares of the population only gradually change. As a
result, immigration issues may be important in a context of relatively low
shares of Turkish and/or Muslim migrants, and not important in a context
of relatively high shares. In this article, we consider immigration issue salience
as a predictor of support for Turkey’s membership along with actual immigration shares. We thus hypothesize the following, in analogy to our hypotheses stated earlier in the text:
FOCUSING ON DIFFERENCES?
485
H4: The greater the share of the population of Turkish descent in a country, the
stronger the negative effect of anti-immigrant attitudes on support for Turkey’s
membership.
Of course, with the same aforementioned arguments, we may argue whether
the effect of anti-immigrant attitudes on support for Turkey’s accession
is affected by the presence of Turkish immigrants, or affected by Muslims,
or affected through indirect contact with immigrants. This leads to the
hypotheses:
H5: The greater the share of the population in a country that is Muslim, the
stronger the negative effect of anti-immigrant attitudes on support for Turkey’s
membership.
H6: The more the immigration issue is important in a country, the stronger
the negative effect of anti-immigrant attitudes on support for Turkey’s
membership.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
certain aspects of an issue and neglecting certain others, a particular interpretation of the issue is promoted above others (e.g., Entman, 1993). For
example, an individual can frame the issue of enlargement in terms of the
economy or in terms of in-group and out-groups. The chosen frame would
lead this individual to interpret enlargement, respectively, as an issue of
weighing gains and losses, or as an issue of whether the applicant country
belongs to ‘‘us.’’ To be able to apply a certain ‘‘frame’’ to a specific issue,
considerations belonging to that particular frame need to be stored in memory to be available to the individual (e.g., Chong & Druckman, 2007b;
Slothuus & De Vreese, 2010). And the presence of Turkish immigrants
makes it more likely that a frame defining Turkish immigrants as an outgroup is available.
In a country where only few Turkish immigrants reside, individuals are
less likely to have thought about Turkish immigrants. But the more Turkish
immigrants in a country, the more likely individuals are to have thought about
them as a group, as it is more likely they encounter these immigrants (either in
real life, through interpersonal communication, or in the media). Individuals
who are negatively predisposed toward immigrants (and who thus, we assume,
have a strong tendency to categorize) are likely to have thought about the
Turkish immigrants as an out-group. Thus, the frame of people of Turkish
descent as an out-group is most available for those individuals who have
strong anti-immigrant attitudes and live in a country where there are relatively
many Turkish immigrants. And when confronted with the question
whether someone supports or opposes Turkey’s EU membership, those who
have the frame of Turkish immigrants as an out-group most readily available
are most likely to use the frame and show a negative bias toward the outgroup, rejecting Turkey’s EU membership. We thus hypothesize the
following:
486
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
Methods
Survey
Variables
Support for Turkey’s membership. The dependent variable is the
degree to which an individual EU citizen supports or opposes Turkey’s membership of the EU. We asked the respondents to score on a scale from 1 to 7
whether they are ‘‘strongly against’’ Turkey’s EU membership (1) or ‘‘strongly
in favor’’ (7) (see Appendix for precise question wording and descriptive
statistics). As we are interested in country difference, we plotted the means
of the dependent variable for each country in Figure 1. In Austria, citizens are
the least supportive of Turkey’s EU membership, with an average score of
2.14; in Poland, citizens are the most supportive of Turkey’s EU membership,
although even here the mean lies below the middle point of the scale.
Anti-immigrant attitudes. The main independent variable is an individual EU citizen’s attitude toward immigrants. As we follow Sniderman et al.
(2000) that the collection of anti-immigrant attitudes is a measure of the
readiness to categorize and subsequently negatively evaluate the ones who
are categorized as different, we do not make a distinction between economic
or cultural threat perceptions. We measured anti-immigrant attitudes using
1
The results presented are unweighted. When applying a weight with regard to age, gender, education,
and region of residence, the results are similar and lead to the same conclusions. Weighted results are
available from the authors on request.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
We conducted a survey in 20 countries of the EU, 3 weeks before the 2009
European Parliament Elections. In selecting which countries to include, we
took into consideration that the sample would include larger and smaller
member states, countries from North, South, East, and West, and long-term
and new members to the EU. The countries included were Austria, Belgium,
Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary,
Italy, Ireland, Latvia, Lithuania, The Netherlands, Poland, Portugal,
Slovakia, Spain, Sweden, and UK. From the TNS databases and their partners, a sample was drawn, with quotas enforced on age, gender, and education
to ensure representativeness.1 A total of 26,344 respondents participated. The
average response rate (AAPOR RR1) was 23% (with a minimum of 13% in
Denmark and a maximum of 46% in Lithuania).
The questionnaire was developed in English and translated by TNS (who
also translated the Eurobarometer surveys) into different languages. As an
additional check, all translated questionnaires were retranslated back into
English. Irregularities and problems arising from this process were resolved
by deliberation.
FOCUSING ON DIFFERENCES?
487
Figure 1
Mean support for Turkey’s EU membership per country
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
five items, which also include the items McLaren used to measure economic
and cultural threat to the in-group. Using exploratory factor analysis (principal
component), we found that the items loaded onto one factor (eigenvalue ¼ 3.12; 62.39% explained variance) and by taking the mean we
constructed a highly reliable scale (Cronbach’s ¼ .84).
Number of Turkish migrants. A first contextual variable is the share of
a country’s population that is of Turkish origin. We retrieved these data using
the Eurostat online database. Concerning the number of Turkish citizens in
2009 in 16 countries, this was unproblematic. For Belgium and Lithuania, the
last estimate was from 2008, for France 2005, and for Greece 2001. For these
countries, the most recent estimate was used.
Number of Muslims. A second contextual variable is the percentage of
Muslims in each country. We used estimates from the report on the global
Muslim population of the Pew Forum on Religion and Public Life (2009).
The sources used in the report include national censuses and general population surveys, such as the European Social Survey and the World Values
Survey. The estimates of the 2009 figures are projections based on data
from 1999 to 2009 under the assumption that the Muslim population would
grow at the same rate as the general population.
Importance of the immigration issue. To assess the degree to which
the immigration issue is important in a polity, we focus on the general public,
on political parties, and on the national media. From various sources we
constructed country-level variables for each of these.
488
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
2
We also ran analyses without weighing for party size in parliament. This leads to similar results,
available from the authors on request.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
For the degree to which the public finds the immigration issue important,
we used the data from our survey. We asked respondents to rate on a scale
from 0 (not at all important) to 6 (very important) how important they thought
the immigration issue was. As we are interested in the degree to which the
issue is important in each country, we aggregated up individual scores by
calculating their country-level means.
We used data from the Chapel Hill Expert Survey that was conducted in
2006 to construct our measure of the importance political parties attach to the
immigration issue. In this survey, professional researchers who have published
on political parties in a particular country (excluding graduate students, party
officials, and journalists) were contacted. From each country, the principle
investigators received between 4 and 13 filled in questionnaires. In the questionnaire, the experts were asked to score political parties on a number of
issues on both the position each party took on the issue and on the importance
it attached to the issue (for a more elaborate discussion of the CHESS data,
see Hooghe et al., 2010). We used the items by which parties were scored on
how important immigration policy, integration of immigrants, and issues relating to ethnic minorities were to them. The items load onto one factor (eigenvalue ¼ 2.41; 80.26% explained variance), and by calculating the mean we
construct a highly reliable scale (Cronbach’s ¼ .88). Country scores were
computed by taking the country mean, weighing each party by its size in
their national parliament.2
To assess the importance of the immigration issue in the national media,
we used the media dataset of the 2009 European Election Study (details about
the media study can be found in: Schuck, Xezonakis, Banducci, & De Vreese,
2010). In this dataset, three national newspapers and two national televisionnews shows were analyzed in the 3 weeks before the 2009 European
Parliament Election. The unit of coding in the media dataset is the news
story, where a story in a newspaper is defined as an article and in television-news defined by its topic. We used items about the primary, secondary,
or tertiary topic of the story to score whether immigration-related issue were
part of the story. For each country, we then calculated what the percentage of
the stories of which the primary, secondary, or tertiary topic was immigration.
With both tabloids and quality newspapers from each country and television
news from both public broadcasting and commercial stations, the chosen outlets in the European Parliament Election Study media dataset are a representative sample of the media in each country.
Controls. In our models, we will control for the most relevant factors
from recent literature and demographics. They include exclusive national
identity (Hooghe & Marks, 2004), government satisfaction (Franklin, Van
489
FOCUSING ON DIFFERENCES?
Table 1
Country Level Correlations
Number of
Number of
Importance
Importance
Turks
Muslims
public
parties
Number of
Muslims
Importance
public
Importance
parties
Media
salience
.67***
.24
.35
.42*
.52**
.66***
.05
.25
.57***
.59***
der Eijk, & Marsh, 1995), economic evaluations (Gabel & Palmer, 1995),
political ideology (Gabel, 1998), political interest, age, gender, education,
and ethnicity.
Data Analysis
Because our dataset consists of individual respondent clustered in different
countries, we first assess the intraclass correlation coefficient. We find a significant proportion of the variation (5.9%) at the country level, and thus use
multilevel modeling techniques (see, e.g., Hox, 2002), also known as hierarchical linear modeling, to test our hypotheses.
We start our analysis with two baseline models: one random-intercept only
model, by which we assess the variation in support for Turkey’s EU membership at the country level, and one which only contains individual-level
predictors, by which we assess the variation in the effect of anti-immigrant
attitudes on such support. Consecutively, we add the country-level variables.
Given the strong correlation between the relative number of Turks and the
relative number of Muslims, and the strong correlations between the various
indicators of importance of the immigration issue (Table 1), it would be
unwise to add all contextual variables in one model. We can, however, combine one of the proportion indicators with one of the issue salience indicators.
First, we test the main effects of the contextual predictors (H1, H2 and H3),
and, second, the interactions of the contextual predictors with anti-immigrant
attitudes (H4, H5 and H6).
Results
We first turn to Model 1 (Table 2). We find an average intercept over all
countries of 3.08. Across the 20 countries in our dataset, we find a variance of
0.220, which significantly differs from zero at the p < .001 level. This means
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Note: Entries are Pearson’s r correlations between independent variables at the country level. N ¼ 20.
****p < .001; ***p < .01; **p < .05; *p < .1 (two-tailed).
490
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
Table 2
Multilevel models explaining support for Turkey’s EU membership
Model 2
3.08**** (.11)
4.22**** (0.10)
0.36**** (0.02)
0.09**** (0.01)
0.08**** (0.01)
0.08**** (0.01)
0.021**** (0.00)
0.00 (0.01)
0.01**** (0.00)
0.12**** (0.02)
0.16**** (0.03)
0.10**** (0.03)
0.15**** (0.05)
0.220****
0.125***
0.010***
3.027****
104,068.46
3.495****
107,815.89
Note: Level 1: individual, N ¼ 26,344; level 2: country, N ¼ 20. Standard errors within parentheses.
****p < .001; ***p < .01 (one-tailed).
that there is cross-country variation in support for Turkey’s membership.
Turning to Model 2 we see that the individual-level predictors do explain
some of the cross country-level variation, but the remaining variance (0.125) is
still significant at the p < .01 level.
Concerning anti-immigrant attitudes, we find in Model 2 of Table 2 a
significant effect (at the p < .001 level) with a coefficient of 0.36. This means
that for every 1 point score more negative toward immigrants (on a scale from
0 to 6), support for Turkey’s membership drops on average 0.36 points (on a
scale from 1 to 7). Again, this is an average effect across the 20 countries in
the data. The variance of the effect is 0.010 and significantly different from
zero at the p < .01 level. Therefore, there is also cross country variation in the
effect of anti-immigrant attitudes on support for Turkey’s membership.
Turning our hypotheses, we should start out by reminding the reader of
the modest size of the country-level variation: it is only 5.9% of the total
variation in the dependent variable (and 4.0% of the variation left after adding
the individual-level predictors). Thus, the effects of country-level factors that
we report in this section are meaningful yet small. In Models 3, 4, and 5
(Table 3), we test the main effect of the presence of Turkish immigrants,
controlled for the three different indicators of importance of immigration. In
all three models, we find a significant negative effect of the relative number of
Turkish immigrants on support for Turkey’s membership (ranging from
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Intercept
Anti-immigrant attitudes
Exclusive national identity
Government satisfaction
Economic evaluations
Political ideology (left-right)
Political interest
Age
Female
Education low
Education high
Member ethnic minority
Variance of random components
Intercept
Anti-immigrant attitudes effect
Residual
Deviance (2LL)
Model 1
Model 4
Model 6
Model 7
Model 8
.180
.007
104,064.86
.128
.003
104,064.90
.078
.004
104,064.70
.095
.005
104,064.21
3.027****
3.027****
0.115***
0.010***
0.137***
0.010***
3.027****
9.36 (9.06)
0.10*** (0.04)
3.027****
0.07 (0.07)
0.09** (0.05)
0.147***
0.010***
0.10** (0.04)
0.15 (0.15)
0.109***
0.010***
10.38 (8.77)
0.36*** (0.14)
0.36**** (0.02) 0.39**** (0.02) 0.36**** (0.02) 0.36**** (0.02)
Model 5
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Note: Level 1: individual, N ¼ 26,344; level 2: country, N ¼ 20. Standard errors within parentheses. Models also include intercept, exclusive national identity, government
satisfaction, economic evaluations, education low, education high, age, gender, member of ethnic minority, political ideology, and political interest, but are not shown in the
table.
****p < .001; ***p < .01; **p < .05.1 (one-tailed).
Anti-immigrant
0.36**** (0.02) 0.36**** (0.02)
attitudes
Number of Turks
0.32** (0.14)
0.33** (0.15)
Number of Muslims
Importance public
0.13 (0.15)
Importance parties
0.07 (0.06)
Media salience
Variance of random components
Intercept
0.110***
0.104***
0.010***
0.010***
Anti-immigrant
attitudes effect
Residual
3.027****
3.027****
Proportions of reduced variances (compared with Model 1)
Intercept
.120
.166
Anti-immigrant
.001
.001
attitudes effect
Deviance (2LL)
104,064.49
104,064.51
Model 3
Table 3
Multilevel models explaining support for Turkey’s EU membership
FOCUSING ON DIFFERENCES?
491
492
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
3
To test whether the difference between the effect sizes is significant we stacked the data. In this new
dataset, each individual is represented twice: once in combination with the number of Turkish immigrants
and once with the number of Muslims. Because of the different range of the presence indicators, we
standardized them both to make them comparable. Consecutively, we analyzed models with an interaction
of the standardized number of Turkish immigrants/Muslims and a variable identifying whether the presence indicator refers to the number of Turkish immigrants or the number of Muslims. This interaction is
not significant at the p ¼ .1 level (two-tailed), which implies that the effect sizes do not differ significantly.
4
To illustrate to the conditionality of the effect of anti-immigrant attitudes on support for Turkey’s EU
membership, graphs are found in the Appendix.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
0.32 in Model 2 to 0.36 in Model 4). The negative coefficient is as expected, as it means that more Turkish immigrants lead to less support for
Turkish immigrants. The observed Turkish immigrant share ranges from
0% to just >2%. Model 2 predicts a difference in support of 0.70 (on a
7-point scale) between the country with the lowest and the country with the
highest Turkish immigrant share. These findings support our first hypothesis.
In Models 6, 7, and 8 (Table 3) our H2, concerning the main effect of the
presence of Muslims, is tested. In these three models, we also find significant
negative effects. With coefficients between 0.09 and 0.10, the effect of the
presence of Muslims seems to be much weaker than the effect of the presence
of Turkish immigrants. However, if we take into consideration that the relative number of Muslims in each country ranges between 0% and almost 6%,
Model 4 estimates a maximum difference of 0.55. This implies that the relative number of Muslims has a weaker effect on support for Turkey’s accession
than the relative number of Turks, but the difference is not as large as a first
look at the coefficients may suggest and is not statistically significant.3
The last hypothesis about main effects was about the importance of the
immigration issue in the polity (H3). In Models 3 and 6, this hypothesis is
tested using importance for citizens; in Models 4 and 7 using importance for
political parties; in Models 5 and 8 using importance in the media. In none of
these models we find significant effects; thus, the data do not support our
hypothesis.
We now turn to the hypotheses about the variation in the effect of antiimmigrant attitudes on support for Turkey’s membership. Models 9, 10, and
11 in Table 4 are used to test H4: The greater the share of the population of
Turkish descent in a country, the stronger the negative effect of anti-immigrant attitudes on support for Turkey’s membership. The interaction between
anti-immigrant attitudes and the relative number of Turkish immigrants has a
significant negative effect in all three models. This negative interaction effect
means that the more people of Turkish descent in a country, the stronger the
effect of anti-immigrant attitudes on support for Turkey’s membership.
Between countries with the lowest and highest number of Turkish immigrants,
the models predict a difference in effect of anti-immigrant attitudes of 0.12 in
Model 10 and 0.21 in Model 11. This supports our fourth hypothesis.4
Model 10
Model 11
Model 12
.120
.540
.059
.569
104,050.99
3.027****
3.027****
104,049.17
0.109***
0.005***
0.03*** (0.01) [.398]
0.06* (0.03) [.138]
0.04 (0.16) [.002]
0.117***
0.004***
5.96*** (1.95) [.356]
0.10*** (0.03) [.360]
0.27**** (0.03) [.870]
.017
.647
104,046.52
104,043.34
3.027****
3.027****
.150
.720
0.122***
0.004***
0.105***
0.003***
5.24*** (1.83) [.334]
0.03**** (0.01) [.476]
0.02** (0.01) [.279]
0.05**** (0.01) [.477]
Model 14
0.25**** (0.03) [.862]
Model 13
0.08 (0.06) [.079]
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Note: Level 1: individual, N ¼ 26,344; level 2: country, N ¼ 20. Standard errors within parentheses. Between the square brackets is the proportion reduced variance in the
effect of anti-immigrant attitudes compared with a model where the specific variable is omitted. Models also include intercept, exclusive national identity, government
satisfaction, economic evaluations, education low, education high, age, gender, member of ethnic minority, political ideology, political interest, and the main effects of the
contextual variables, but are not shown in the table.
****p < .001; ***p < .01; **p < .05; *p < .1 (one-tailed).
Anti-immigrant attitudes
0.02 (0.17) [.000]
0.07 (0.07) [.058]
(AIA)
AIA* number of Turks
0.09*** (0.04) [.291]
0.06** (0.03) [.182]
AIA* number of Muslims
AIA* importance public
0.07** (0.04) [.194]
AIA* importance parties
0.05**** (0.01) [.529]
AIA* media salience
Variance of random components
Intercept
0.106***
0.102***
Anti-immigrant attitudes
0.005***
0.003***
effect
Residual
3.027****
3.027****
Proportions of reduced variances (compared with Model 1)
Intercept
.147
.179
Anti-immigrant attitudes
.457
.682
effect
Deviance (-2LL)
104,053.62
104,044.91
Model 9
Table 4
Multilevel models explaining support for Turkey’s EU membership
FOCUSING ON DIFFERENCES?
493
494
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
Conclusion
In this article, we asked the question which contextual conditions affect levels
of support for Turkey’s EU membership. Following the work of McLaren
(2007) and building on conflict theory, we first hypothesized that in a country
with more Turkish immigrants support for Turkey’s accession would be
lower. We went beyond the literature in two ways. First, as Turkish
5
We tested this in a similar fashion as we tested the difference between the main effects. Because we are
interested in whether sizes of interaction effects differ, we tested this with a three-way interaction with the
identifier variable.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
For H5, about the moderation by the relative number of Muslims, and
H6, about the moderation by the importance of the immigration issue, we find
similar results. In all models (Models 12, 13, and 14 for the relative number of
Muslims, Models 9 and 12 for importance for the public, Models 10 and 13
for importance of political parties, and Models 11 and 14 for media salience),
we find that the effect of anti-immigrant attitudes is significantly moderated
by our contextual variables. Between low and high score countries on these
contextual variables, we find differences in effects of anti-immigrant attitudes
on support for Turkey’s membership between 0.12 (for the relative number of
Muslims in Model 12) and 0.25 (for importance for political parties in
Model 9). These findings support the fifth and the sixth hypothesis.
The question that remains is which contextual variables best explain the
variation in strength of the effect. For this, we also calculated for each interaction how great a proportion of the variation could not be explained when the
interaction was omitted from the model. With this, we can see how much of
the variation in the effect is explained by this interaction. These numbers are
represented in Table 4 between the square brackets. Here, we see that the
models predict that the relative number of Muslims explains more variation in
the effect of anti-immigrant attitudes than the relative number of Turkish
immigrants does. This difference is, however, not statistically significant.5
Comparing the presence of Turkish immigrants and/or Muslims with the
importance of the immigration issue, which contextual factor is most important depends on which indicator we use. Importance for political parties is the
strongest predictor of the effect of anti-immigrant attitudes. Importance for
citizens is a significant predictor, but not as strong as the relative number of
Turkish immigrants or Muslims. But as all the interaction effects hold in all
models, importance of the immigration issue in the polity and number of
Turkish immigrant or Muslims each predict substantial variation in the
strength of the effect of anti-immigrant attitudes on support for Turkey’s
membership.
FOCUSING ON DIFFERENCES?
495
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
immigrants may be primarily seen as Muslims and Turkey may be primarily
seen as a Muslim country, we also hypothesized that more Muslims would
lead to less support for Turkey’s membership. Second, as most citizens do not
have direct experience with Turkish immigrants or Muslims, we argued that
what is crucial is their indirect experience with immigrants of hearing about
immigrants. We expected that when the immigration issue is more important
in a polity, citizens would have more indirect experience with immigrants, and
thus hypothesized that the more important the immigration issue, the less
support we find for Turkey’s EU membership.
As our expectations with regard to the effects of the contextual conditions
is based on Conflict Theory, and thus the expectation that citizens define
Turkish immigrants as an out-group, we also looked at how these conditions
moderate the effect of anti-immigrant attitudes on Turkey’s EU membership
(because the effect of anti-immigrant attitudes on support is also based on
defining others as an out-group). We hypothesized that the higher the relative
number of people of Turkish descent, the higher the relative number of
Muslims, and the more important the immigration issue is in a country, the
stronger the effect of anti-immigrant attitudes on support for Turkey’s EU
membership.
Using data from a survey of EU citizens in 20 countries, we found partial
support for our hypotheses concerning the main effects. Keeping in mind that
the cross-country variation was low but significant, we found that the relative
number of Turks and the relative number of Muslims turn out to matter, with
no significant difference between the effect sizes of both indicators. We found
no support for our expectation that the importance of the immigration issue
affected support for Turkey’s EU membership. We expected the main effect
of population proportion Muslims to have a stronger effect than population
proportion of Turkish descent. That we do not find this proves us wrong
concerning our reasoning that Muslims might be more visually distinctive than
Turks and thus show greater contact effects.
The hypotheses about the moderation of the effect of anti-immigrant attitudes were all supported by our data. Similar as for the main effects, there
was no significant difference between the number of Turkish immigrants and
the number of Muslims in predicting the effect size of anti-immigrant attitudes on support for Turkey’s EU membership. The importance of the immigration issue is a significant predictor of the effect size of anti-immigrant
attitudes. When operationalized as the importance political parties attach to the
immigration issue, it is also the strongest predictor. With the contextual predictors affecting the effect size of anti-immigrant attitudes, it seems that the
relevance of the contextual factors for support for Turkey’s EU membership is
primarily indirect through affecting the strongest predictor of support in the
literature.
496
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Theoretically, we argued that the presence of Turkish immigrants
(or Muslims) moderates the effect of anti-immigrant attitudes on support
for Turkey’s EU membership. But could this moderation also work the
other way around? Could it be the case that individuals with strong antiimmigrant attitudes react more strongly to the presence of Turkish immigrants? This may seem likely, and with our data, we are unable to test for
the direction of the moderation. We argue, however, that for the interaction
between anti-immigrant attitudes and ‘‘indirect contact’’ there are theoretical
reasons to expect the direction we proposed in this article. We argued that
through indirect contact citizens learn to think in terms of ‘‘us’’ and ‘‘them,’’
and thus when the issue of Turkey’s potential accession comes to mind, it is
likely that with more indirect contact, citizens think more in terms of ‘‘us’’
versus ‘‘them,’’ and the effect is thus stronger. In this case, the other direction
is less likely: Why would individuals who think more in terms of ‘‘us’’ and
‘‘them’’ be more opposed to Turkey’s EU membership, if the immigration
issue is highly politicized? As we contrast ‘‘direct’’ and ‘‘indirect’’ contact,
and argue that both types of contact effect support for Turkey’s membership through the same mechanism, we argue a similar direction of the
moderation.
Our findings on the main effects imply that public opinion is affected by
direct experiences with ‘‘others,’’ and not by indirect experience through the
importance of the immigration issue. We should, however, be careful with
how to interpret this direct experience. As we argued earlier in the text, the
presence of Turkish immigrants or Muslims does not imply direct contact at
the personal level. Although given our method of analysis, we can strictly
speaking only conclude contact at the country level, we may argue that citizens
are aware of the presence of either Turkish immigrants or Muslims through
which their presence works down to the personal level. How this contact at
the country level works down to the personal level should, of course, be
examined in future studies.
The significant effect of the relative number of Muslims indicates that for
citizens the religion of Turks is of importance in deciding on support for
Turkey’s membership, which is in line with other studies (Azrout, Van
Spanje, & De Vreese, 2013; Hobolt, Van der Brug, De Vreese, Boomgaarden,
& Hinrichsen, 2011; Van Spanje et al., 2010). However, given the similar effect
size of the relative number of Turkish immigrants and the relative number of
Muslims, we may wonder whether citizens actually differentiate between Turks
and Muslims when forming an opinion about the potential accession of a
Muslim country. In line with this, we already noted that several authors claim
that Turks are increasingly being stereotyped as Muslims (e.g., Marranci, 2004;
Poynting & Mason, 2007). But if citizens have trouble to differentiate between
FOCUSING ON DIFFERENCES?
497
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Turks and Muslims, does that mean that the presence of Turkish immigrants
also affects support for other Muslim countries aspiring EU membership?
Also, if the presence of Muslims is important in making people think in
terms of ‘‘us’’ and ‘‘them’’ (i.e., affects the effect of anti-immigrant attitudes
on support for membership), this raises the question to what degree these
findings may hold for other, non-Muslim, applicant countries. We argued that
the presence of Muslims increases the likelihood of citizens having thought
about them, and for those who hold strong anti-immigrant attitudes it increases the likelihood to have thought about Muslims as an out-group. When
the consideration of Muslims as an out-group is stored in memory and thus
available, it is likely to be applied when asked about a Muslim country.
But except for a consideration to be available, it is described in the framing
literature that to be applied a frame should also be applicable (e.g., Chong &
Druckman, 2007a; Scheufele & Tewksbury, 2007). A frame where Muslims
are described as ‘‘others’’ is not likely to be applicable for non-Muslim countries. But the question is whether citizens learn to focus on differences with
Muslims or to focus on differences in general from the presence of Muslims in
their country. Of course, considerations of focusing on differences in general
can indeed be applied to non-Muslim countries. Thus, future studies should
examine whether the considerations to focus on differences, learned from
the presence of Muslims, are Muslims specific or of a more general nature.
Finally, we found that mediated contact through the importance of the
immigration issue explained part of the variation in the strength of the effect
of anti-immigrant attitudes on support for Turkey’s EU membership.
Although all our importance indicators had a significant effect, the strength
of the effects clearly differed. Importance for the public was the weakest
predictor of the effect size, salience in the media of the immigration issue
being stronger, and importance of political parties being the strongest predictor. This implies that citizens are probably less aware of what other people
think about the immigration issue and more susceptible to cues from the
media and politicians. Of course, we should ask ourselves whether cues
from the media and from politicians are not themselves cued by the public,
i.e. a representation of ideas that exist among the public. Steenbergen et al.
(2007) showed that political elites both respond to and shape ideas of their
constituencies, thus making politicians (at least partially) instigators. Also,
politicians because of their authoritative status, often act as primary definers
of issues and ideas in the media (Benson, 2004; Hall, Critcher, Jefferson,
Clarke, & Roberts, 1978). In light of the position politicians take in cueing
both public and media, our findings may not imply direct influence of political
elite’s on citizens’ thoughts, it does seem to imply that elites have the ability to
steer the criteria by which citizens form their opinion.
498
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
References
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Allport, G. W. (1954). The nature of prejudice. London: Addison-Wesley.
Amir, Y. (1976). The role of intergroup contact in change of prejudice and race
relations. In P. A. Katz (Ed.), Towards the elimination of racism. (pp. 245–308).
New York: Pergamon Press.
Anderson, C. J., & Reichert, M. S. (1996). Economic benefits and support for membership in the EU: A cross-national analysis. Journal of Public Policy, 15, 231–249.
doi:10.1017/S0143814X00010035.
Azrout, R., Van Spanje, J. H. P., & De Vreese, C. H. (2011). Talking
turkey: Anti-immigrant attitudes and their effect on support for Turkish membership of the EU. European Union Politics, 12, 3–19. doi:10.1177/
1465116510389498.
Azrout, R., Van Spanje, J. H. P., & De Vreese, C. H. (2013). A threat called turkey:
Perceived religious threat and support for EU entry of Croatia, Switzerland and
Turkey. Acta Politica, 48, 2–21. doi:10.1057/ap.2012.20.
Benson, R. (2004). Bringing the sociology of media back in. Political Communication,
21, 275–292. doi:10.1080/10584600490481299.
Blalock, H. M. (1967). Towards a theory of minority-group relations. New York:
Capricorn Books.
Bobo, L. (1988). Group conflict, prejudice, and the paradox of contemporary racial
attitudes. In P. A. Katz & D. A. Taylor (Eds.), Eliminating racism: Profiles in
controversy. (pp. 85–114). New York: Plenum Press.
Chong, D., & Druckman, J. N. (2007a). Framing theory. Annual Review of Political
Science, 10, 103–26. doi:10.1146/annurev.polisci.10.072805.103054.
Chong, D., & Druckman, J. N. (2007b). A theory of framing and opinion formation in
competitive elite environments. Journal of Communication, 57, 99–118. doi:10.1111/
j.1460-2466.2006.00331.x.
De Vreese, C. H., & Boomgaarden, H. G. (2005). Projecting EU referendums: Fear of
immigration and support for European integration. European Union Politics, 6,
59–82. doi:10.1177/1465116505049608.
De Vreese, C. H., Boomgaarden, H. G., & Semetko, H. A. (2008). Hard and soft:
Public support for turkish membership in the EU. European Union Politics, 9,
537–556. doi:10.1177/1465116508095149.
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal
of Communication, 43, 51–58. doi:10.1111/j.1460-2466.1993.tb01304.x.
Forbes, H. D. (1997). Ethnic Conflict: Commerce, Culture, and the Contact Hypothesis.
New Haven, CT: Yale University Press.
Franklin, M., Van der Eijk, C., & Marsh, M. (1995). Referendum outcomes and trust
in government: Public support for Europe in the wake of Maastricht. West European
Politics, 18, 101–117. doi:10.1080/01402389508425093.
Gabel, M. (1998). Public support for European integration: An empirical test of five
theories. The Journal of Politics, 60, 333–354. doi:10.2307/2647912.
Gabel, M., & Palmer, H. (1995). Understanding variation in public support for
European integration. European Journal of Political Research, 27, 3–19.
doi:10.1111/j.1475-6765.1995.tb00627.x.
FOCUSING ON DIFFERENCES?
499
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Garry, J., & Tilley, J. (2009). The macroeconomic factors conditioning the impact of
identity on attitudes towards the EU. European Union Politics, 10, 361–379.
doi:10.1177/1465116509337829.
Gerhards, J., & Hans, S. (2011). Why not turkey? attitudes towards
turkish membership in the EU among citizens in 27 European countries.
Journal of Common Market Studies, 49, 741–766. doi:10.1111/j.14685965.2010.02155.x.
Giles, M. W., & Hertz, K. (1994). Racial threat and partisan identification. American
Political Science Review, 88, 317–326. doi:10.2307/2944706.
Hall, S., Critcher, C., Jefferson, T., Clarke, T., & Roberts, B. (1978). Policing the crisis
mugging, the state, and law and order. New York: Holmes & Meier.
Hobolt, S. B., Van der Brug, W., De Vreese, C. H., Boomgaarden, H. G., &
Hinrichsen, M. C. (2011). Religious intolerance and euroscepticism. European
Union Politics, 12, 359–379. doi:10.1177/1465116511404620.
Hooghe, L., Bakker, R., Brigevich, A., De Vries, C. E., Edwards, E. E., Marks, G.,
Rovny, J., Steenbergen, M., & Vachudova, M. (2010). Research note: Reliability
and validity of the 2002 and 2006 chapel hill expert surveys on party positioning.
European Journal of Political Research, 49, 687–703. doi:10.1111/j.14756765.2009.01912.x.
Hooghe, L., & Marks, G. (2004). Does identity or economic rationality drive public
opinion on European integration? PS: Political Science and Politics, 37, 415–420.
doi:10.2307/4488854.
Hox, J. (2002). Multilevel analysis. techniques and applications. Mahwah, NJ: Lawrence
Erlbaum.
Jones, E., & Bijl, N. V. D. (2004). Public opinion and enlargement: A
gravity approach. European Union Politics, 5, 331–51. doi:10.1177/
1465116504045156.
Karp, J. A., & Bowler, S. (2006). Broadening and deepening or broadening versus
deepening - the question of enlargement and Europe’s ‘hesitant Europeans’.
European Journal of Political Research, 45, 369–90. doi:10.1111/j.14756765.2006.00302.x.
Marranci, G. (2004). Multiculturalism, Islam and the clash of civilizations theory:
Rethinking Islam phobia. Culture and Religion, 5, 105–117. doi:10.1080/
0143830042000200373.
McLaren, L. M. (2002). Public support for the European union: Cost/Benefit analysis
or perceived cultural threat? The Journal of Politics, 64, 551–66. doi:10.1111/14682508.00139.
McLaren, L. M. (2007). Explaining opposition to turkish membership of the EU.
European Union Politics, 8, 251–278. doi:10.1177/1465116507076432.
Oliver, J. E., & Mendelberg, T. (2000). Reconsidering the environmental determinants of white racial attitudes. American Journal of Political Science, 44, 574–589.
doi:10.2307/2669265.
Paolini, S., Harwood, J., & Rubin, M. (2010). Negative intergroup contact makes
group membership salient: Explaining why intergroup conflict endures.
Personality and Social Psychology Bulletin, 36, 1723–1738. doi:10.1177/
0146167210388667.
500
INTERNATIONAL JOURNAL OF PUBLIC OPINION RESEARCH
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Pettigrew, T. F., & Tropp, L. R. (2006). A meta-analytic test of intergroup contact
theory. Journal of Personality and Social Psychology, 90, 751–783. doi:10.1037/00223514.90.5.751.
Pew Forum on Religion & Public Life. (2009). Mapping the Global Muslim Population.
A Report on the Size and Distribution of the World’s Muslim Population. Washington
DC: The Pew Research Centre.
Poynting, S., & Mason, V. (2007). The resistible rise of Islam phobia: Anti-Muslim
racism in the UK and Australia before 11 September 2001. Journal of Sociology, 43,
61–86. doi:10.1177/1440783307073935.
Putnam, R. D. (2007). E pluribus unum: Diversity and community in the twenty-first
century. the 2006 Johan skitter prize lecture. Scandinavian Political Studies, 30,
137–174. doi:10.1111/j.1467-9477.2007.00176.x.
Quillian, L. (1995). Prejudice as a response to perceived group threat: Population
composition and anti-immigrant and racial prejudice in Europe. American
Sociological Review, 60, 586–611. doi:10.2307/2096296.
Scheufele, D. A. (1999). Framing as a theory of media effects. Journal of
Communication, 49, 103–122. doi:10.1111/j.1460-2466.1999.tb02784.x.
Scheufele, D. A., & Tewksbury, D. (2007). Framing, agenda setting, and priming:
The evolution of three media effects models. Journal of Communication, 57, 9–20.
doi:10.1111/j.0021-9916.2007.00326.x.
Schuck, A. R. T., Xezonakis, G., Banducci, S. A., & De Vreese, C. H. (2010). EES
(2009) media study data advance release documentation, 31/03/2010 (www.piredeu.eu)
. Unpublished manuscript.
Slothuus, R., & De Vreese, C. H. (2010). Political parties, motivated reasoning, and
issue framing effects. Journal of Politics, 72, 630–45. doi:10.1017/
S002238161000006X.
Sniderman, P. M., Hagendoorn, L., & Prior, M. (2004). Predisposing factors and
situational triggers: Exclusionary reactions to immigrant minorities. American
Political Science Review, 98, 35–49. doi:10.2307/4145295.
Sniderman, P. M., Peri, P., De Figueiredo, R. J. P., & Piazza, T. (2000). The outsider:
Prejudice and politics in Italy. Princeton, NJ: Princeton University Press.
Steenbergen, M. R., Edwards, E. E., & de Vries, C. E. (2007). Who’s cueing whom?
mass-elite linkages and the future of European integration. European Union Politics,
8, 13–35. doi:10.1177/1465116507073284.
Taylor, M. C. (1998). How white attitudes vary with the racial composition of local
populations: Numbers count. American Sociological Review, 63, 512–535.
doi:10.2307/2657265.
Valenty, L. O., & Sylvia, R. D. (2004). Thresholds for tolerance: The impact of racial
and ethnic population composition on the vote for California propositions 187 and
209. Social Science Journal, 41, 433–446. doi:10.1016/j.soscij.2004.04.008.
Van Spanje, J. H. P., Azrout, R., & De Vreese, C. H. (2010). On members and
Muslims: Perceived Islam threat and opinion formation on Muslim countries’ EU
entry. Paper Presented at the Etmaal Voor De Communicatiewetenschap; 5
Februari, Gent, Belgium.
FOCUSING ON DIFFERENCES?
501
Biographical Notes
Rachid Azrout is a Postdoctoral Research Fellow in political communication at the
Amsterdam School of Communication Research, University of Amsterdam.
His research interests include framing effects and public opinion toward the
European Union, with a special interest toward EU enlargement.
Claes de Vreese is a Professor and chair of political communication, University of
Amsterdam. He has published >100 articles in academic journals, and his research is
funded by several national and international grants. His research focuses on political
communication, journalism, direct democracy, and public opinion. For more information see www.claesdevreese.com.
Downloaded from http://ijpor.oxfordjournals.org/ at Universiteit van Amsterdam on December 5, 2013
Joost van Spanje is an Assistant Professor of political communication and quantitative
methodology, University of Amsterdam. He won the 2010 Dutch Political Science
Association (NKWP) Dissertation Award and a 2012 Dutch Science Foundation
(NWO) 3-year ‘‘Veni’’ research grant. His research interests include political behavior, electoral studies, public opinion, and political communication. Joost’s work has
been published in 19 articles in ISI-ranked international journals.